Abstract
A method for verifying a periodic useful signal component of a sensor signal of a flowmeter. To distinguish a useful signal from a noise signal, a periodic measured variable is detected over a period, and a periodic sensor signal is output, the sensor signal is subjected to a time/frequency analysis, and a time-dependent frequency spectrum of the sensor signal is determined, the time-dependent frequency spectrum is examined for a characteristic feature of the useful signal, and if the expected characteristic feature is found in the time-dependent frequency spectrum, the portion of the sensor signal having the characteristic feature is verified as the useful signal component and a flow measured value is determined on the basis of the verified useful signal component, or if the expected characteristic feature is not found, the portion of the sensor signal which does not have the characteristic feature is rejected as the noise signal component.
Claims
1. Method for verifying a periodic useful signal component of a sensor signal of a flowmeter, wherein the useful signal component of the sensor signal is used to determine a flow measured value of a flowing medium, comprising a method cycle of: a detection step of detecting a periodic measured variable over a period t.sub.1, and outputting the periodic sensor signal detected, an analysis step of subjecting the sensor signal to a time/frequency analysis, and determining a time-dependent frequency spectrum of the sensor signal, an examination step of examining the time-dependent frequency spectrum for an expected characteristic feature which is characteristic of a useful signal, and a verification step of verifying a portion of the sensor signal having the characteristic feature as the useful signal component if the expected characteristic feature is found in the time-dependent frequency spectrum and determining a measured flow value based on the verified useful signal component, or a rejection step of rejecting a portion of the sensor signal which does not have the characteristic feature as a noise signal component if the expected characteristic feature is not found in the time-dependent frequency spectrum.
2. Method according to claim 1, further comprising using a short-time Fourier transform for carrying out the time/frequency analysis.
3. Method according to claim 1, wherein the time/frequency analysis is carried out by means of one of a wavelet transform, a Gabor transform or a Hilbert-Huang transform.
4. Method according to claim 1, wherein at least one of a predetermined amplitude modulation or a predetermined frequency modulation of a signal component is selected as the characteristic feature.
5. Method according to claim 1, wherein at least one of a predetermined value for a ratio of amplitudes of a signal component of a fundamental frequency and a signal component of a higher order in a time-dependent frequency spectrum is selected as the characteristic feature or a value for the ratio of the amplitudes of two signal components of higher orders is selected as the characteristic feature.
6. Method according to claim 1, wherein a jitter in the time-dependent frequency spectrum is selected as a characteristic feature.
7. Method according to claim 1, further comprising selecting a subsection of the time-dependent frequency spectrum in a subsection selection step to identify a characteristic feature, and subjecting the subsection to a further time/frequency analysis in a subsection analysis step.
8. Method according to claim 1, further comprising performing said method cycle at least twice chronologically in succession, wherein, after each method cycle, information is stored about the portion of the sensor signal verified as a useful signal component, and after a first method cycle, after each next method cycle, a current, verified useful signal component is compared to the useful signal component of the preceding method cycle, and in the event of a deviation of the compared useful signal components from one another beyond a predetermined tolerance range, a notification is output.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0031] FIG. 1 is a flow chart of a first embodiment of the method according to the invention,
[0032] FIG. 2 is a flow chart of an embodiment for finding characteristic features in a spectrum,
[0033] FIG. 3 is a flow chart of a second embodiment of the method according to the invention,
[0034] FIG. 4a is a plot of a pure useful signal,
[0035] FIG. 4b is a plot of a sensor signal with a useful signal component and a noise signal component,
[0036] FIG. 5 is a spectrogram of a useful signal stable in frequency and amplitude over time,
[0037] FIG. 6 is a spectrogram of an amplitude-modulated useful signal,
[0038] FIG. 7 is a spectrogram of a frequency-modulated useful signal,
[0039] FIG. 8 is a spectrogram of a useful signal modulated by both amplitude and frequency,
[0040] FIG. 9 is a spectrogram of a useful signal with harmonics,
[0041] FIG. 10 is a spectrogram of a useful signal with amplitude-modulated harmonics,
[0042] FIG. 11 is a spectrogram of a useful signal with temporally limited noise signal components and
[0043] FIG. 12 is a spectrogram of a useful signal at a flow pulse.
BRIEF DESCRIPTION OF THE DRAWINGS
[0044] FIG. 1 shows a block diagram of a first embodiment of the method according to the invention for verifying a useful signal component of a periodic sensor signal of a flowmeter. In a detection step 101, a measured variable is detected by a sensor of the flowmeter. The measured variable is detected over a period t.sub.1, wherein the period t.sub.1 can, for example, be specified by a user. The sensor then outputs a periodic sensor signal. The sensor signal has a useful signal component and a noise signal component. In an analysis step 102, the sensor signal is subjected to a time/frequency analysis. The time/frequency analysis is carried out in the embodiment shown by means of a Garbor transform. Not shown, but also covered by the invention is the carrying out of the time/frequency analysis with any other transform, such as a wavelet transform. After the time/frequency analysis, a time-dependent frequency spectrum of the sensor signal is determined. In an examination step 103, the frequency spectrum is examined for one or more characteristic features of the useful signal. If one or more expected characteristic features is found, the portion of the sensor signal having the characteristic features is verified as the useful signal component in a verification step 104. A determination of the flow measured value can then be carried out on the basis of this portion of the sensor signal verified as the useful signal component. If the expected characteristic feature is not found, or if the expected characteristic feature is not found in the spectrum, the corresponding sensor signal component is discarded as a noise signal component in a rejection step 105.
[0045] FIG. 2 shows a block diagram of an embodiment for finding characteristic signals in the frequency spectrum, i.e., it shows an embodiment of the examination step 103. To find expected characteristic features, a subsection of the time-dependent frequency spectrum is selected in a subsection selection step 103a, which is to be examined more closely. The selection of the subsection can be carried out by a user. In a subsection analysis step 103b, the selected subsection is subjected to a further time/frequency analysis. In the example shown, the further time/frequency analysis is carried out by means of a wavelet transform.
[0046] FIG. 3 shows a block diagram of another preferred embodiment of the method for verifying a useful signal component of a periodic sensor signal. The first method steps correspond to the method steps shown in FIG. 1: in a detection step 101, a periodic measured variable is detected over a period t.sub.1 and a periodic sensor signal is output from a sensor. In an analysis step 102, the sensor signal is subjected to a time/frequency analysis. After determining a time-dependent frequency spectrum of the sensor signal, the determined frequency spectrum is examined for a characteristic feature of the useful signal in an examination step 103. In a verification step 104, when the expected characteristic feature is found, the portion of the sensor signal with the characteristic feature is verified as the useful signal component and a flow measured value is determined on the basis of the verified useful signal component. Information about the verified useful signal component is then stored in an information storage step 106. The information can, for example, be the characteristic features or can be the flow measured value determined on the basis of the verified useful signal component. The method described in FIG. 3 is now wherein the method described so far is repeated, i.e., the method steps are carried out several times in succession. In a second detection step 101, a measured variable is now detected by a sensor over a further period t.sub.2 and a periodic sensor signal is output. In a second analysis step 102, the sensor signal is subjected to a time/frequency analysis and a time-dependent frequency spectrum of the sensor signal is determined. In a second examination step 103, the frequency spectrum is examined for a characteristic feature. Here the characteristic feature corresponds to the characteristic feature already sought in the first method cycle. When the expected characteristic feature is found, the sensor signal component with the characteristic feature is verified as the useful signal component in a second verification step 104 and a further flow measured value is determined on the basis of the second verified useful signal component. In a second information storage step 106, the same information concerning the useful signal component is stored as in the first method cycle. In a subsequent comparison step 107, the useful signal component of the second method cycle is compared to the useful signal component of the first method cycle. The comparison is made on the basis of the stored information of the respective useful signal components. If the useful signal components deviate from one another beyond a specified tolerance range, a corresponding notification is output in a notification step 108. The embodiment of the method thus allows a good possibility for monitoring the flow measurement. In particular, long-term monitoring can be implemented with which changes to the flowmeter can be easily detected.
[0047] FIG. 4a shows a spectrum of an undisturbed sensor signal at a time t, i.e., a sensor signal which consists only of a useful signal. The useful signal has an amplitude peak at a fundamental frequency f.sub.0 with the amplitude A.sub.0, a first-order amplitude peak at the frequency f.sub.1 with the amplitude A.sub.1 and a second-order amplitude peak at the frequency f.sub.2 with the amplitude A.sub.2. A characteristic feature of the useful signal shown is that the amplitude A.sub.1 of the first-order amplitude peak is half as large as the amplitude A.sub.0 of the second-order amplitude peak at the fundamental frequency f.sub.0 and that the amplitude A.sub.2 of the second-order amplitude peak is half as large as the amplitude A.sub.1 of the first-order amplitude peak and thus a quarter as large as the amplitude A.sub.0 of the amplitude peak at the fundamental frequency f.sub.0. The characteristic feature chosen here is that the value for the ratio of the amplitudes of the amplitude peak of order O and the amplitude peak of order (O+1) is one half.
[0048] FIG. 4b shows a spectrum at a time t of a sensor signal, which in addition to a useful signal component also has a noise signal component. The useful signal component corresponds to the useful signal shown in FIG. 4a. The noise signal component consists of an additional amplitude peak at the frequency f.sub.s0 with the amplitude A.sub.s0 and an additional amplitude peak at the frequency f.sub.s1 with the amplitude A.sub.s1. The noise signal component can be identified as the noise signal component in that the amplitude ratio A.sub.s1/A.sub.s0 of the amplitude peaks at the frequencies f.sub.s0 and f.sub.s1 is not one half and thus does not exhibit the characteristic feature.
[0049] FIGS. 5 to 12 show time-dependent frequency spectra in the form of spectrograms, each of which shows useful signals with different characteristic features.
[0050] In FIGS. 5 to 12, the time t is plotted on the x-axis, the frequency f is plotted on the y-axis and the amplitude A is plotted in color-coded form. The corresponding color scale is shown to the right of the spectrogram.
[0051] FIG. 5 shows a useful signal stable in frequency and amplitude over time without special characteristic features.
[0052] In order for a signal to be verified as a useful signal, it can be provided that the signal has exactly one characteristic feature. For example, this can be a certain amplitude modulation or a certain frequency modulation. The useful signal shown in FIG. 6 has an amplitude modulation that corresponds to the amplitude modulation defined as a characteristic feature.
[0053] The useful signal shown in FIG. 7 is frequency modulated in a characteristic manner. Thus, if a signal has the corresponding frequency modulation, it is verified as a useful signal.
[0054] In order for a signal to be verified as a useful signal, it can also be provided that it has several characteristic features, for example, that it is amplitude modulated in a certain way and frequency modulated in a certain way. If the signal has only one of the two characteristic features, it is not verified as a useful signal but rejected as a noise signal. The useful signal shown in FIG. 8 has two characteristic features: On the one hand, it has an amplitude modulation defined as a characteristic feature, on the other hand, it also has a frequency modulation defined as a characteristic feature.
[0055] FIG. 9, like FIG. 4a, shows a useful signal with a signal component at a fundamental frequency f.sub.0 with the amplitude A.sub.0, a signal component of the first order at the frequency f.sub.1 with the amplitude A.sub.1 and a signal component of the second order at the frequency f.sub.2 with the amplitude A.sub.2. In contrast to FIG. 4a, however, not only one time point t is shown, but rather the entire spectrogram with its temporal course. It is characteristic of the useful signal shown that the amplitude A.sub.1 of the first-order amplitude peak is in a certain ratio to the amplitude A.sub.0 of the second-order amplitude peak at the fundamental frequency f.sub.0 and that the amplitude A.sub.2 of the second-order amplitude peak is in a certain ratio to the amplitude A.sub.1 of the first-order amplitude peak. The characteristic feature selected here is a certain value for the ratio of the amplitudes of the amplitude peak of order O and the amplitude peak of order (O+1).
[0056] The spectrogram shown in FIG. 10 shows the characteristic feature of the amplitude modulation of the harmonics in addition to the characteristic feature already shown in FIG. 9. A signal is thus verified as a useful signal if, in addition to the given amplitude ratios of the signal components of the various orders, it also has the characteristic amplitude modulation.
[0057] FIG. 11 shows the useful signal shown in FIG. 5 with constant amplitude and constant frequency, wherein the spectrogram shown in FIG. 11 also shows temporally limited noise signal components. The temporally limited noise signals can, for example, result from gas bubbles occurring in the medium if the spectrogram is based on a vortex flow measurement.
[0058] FIG. 12 shows another spectrogram showing a useful signal at a flow pulse. The frequency and amplitude of the signal are characteristically linked.
